Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Emerge - the future?
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5  Next  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
far
Guru
Guru


Joined: 10 Mar 2003
Posts: 394
Location: Stockholm, Sweden

PostPosted: Thu Aug 14, 2003 10:49 pm    Post subject: Reply with quote

Tronic wrote:
But if I had that database in one big file, with an average of 500 chars information per package, you'd still have to scan thru several megs of text and naturally also handle the database at the same time. Potentially you'd have to parse it in UTF-8 too. I don't have any benchmarks here on Python, but that might still be too slow (is it?)

I don't have any benchmarks, but I think execution speed is insignificant next to disk access time. Besides, many of the built-in and library functions in Python are in fact written in C.

Tronic wrote:
Quote:
There is nothing wrong with using xml. Unlike binary data, it is readable (and debuggable) by human beings.


But at the same time it is difficult to read for machines (and it's software that should ever be reading or writing it anyway). The things we are talking about here are simplicity of implementation (escaping all data that isn't ASCII text, escaping XML reserved characters, parsing of tags that use loose syntax (the number of spaces between attributes and other such small things)), storage efficiency (big and ugly tags versus simple binary ones) and performance.

Most of the above is a non-issue. Much of the point with XML is that the implementation is not a problem. You don't have to write your own parser or XML generator, because you can use existing XML libraries and concentrate on the interesting parts.

Using a binary format to save space and time is a silly idea. Linux doesn't run on the VIC-20 so you usually don't have to worry about running out of memory. Unix people tend to hate file formats they can't read with their favourite text editor. And so they should.

Remember: "Premature optimization is the root of all evil."
_________________
The Porthole Portage Frontend
Back to top
View user's profile Send private message
far
Guru
Guru


Joined: 10 Mar 2003
Posts: 394
Location: Stockholm, Sweden

PostPosted: Thu Aug 14, 2003 10:52 pm    Post subject: Reply with quote

Tronic wrote:
While the progress bars can't work when only building a single package, they'd surely be very useful in those operations which take the most time - those which recompile half of the entire system.

(and it doesn't really have to be a bar, many other ways of displaying the information might actually be better)

It does already have such a feature. The title bar of your terminal window should show "compiling package 3 of 7" or something like that.
_________________
The Porthole Portage Frontend
Back to top
View user's profile Send private message
Senso
Apprentice
Apprentice


Joined: 17 Jun 2003
Posts: 250
Location: Montreal, Quebec

PostPosted: Fri Aug 15, 2003 12:01 am    Post subject: Reply with quote

Tronic wrote:

Someone suggested SQL.. Dunno about its search performance either, could be fast too..


It IS extremely fast, compared to normal text searching/indexing. :) It's in binary so not human-readable but MySQL or PostgreSQL (and I hope SQLite) allow a lot of flexibility if you need to change something. Also, Python has a good MySQL support.
Back to top
View user's profile Send private message
NiXZe
Tux's lil' helper
Tux's lil' helper


Joined: 22 May 2003
Posts: 75
Location: Uppsala Sweden

PostPosted: Fri Aug 15, 2003 12:34 am    Post subject: Reply with quote

I've been thinking the system of rsync seams a litle bit od to me or more correctly the use of the ebuilds in the rsync process... (i don't use all the ebuild files why wast my "pressius" hd space) what if instead something like a text file with all .ebuilds availeble when u want to install a package u do a emerge some_package then it downloads a tgz/tar.bz2 with the ebuild and Manifest * * (files neaded) and then goes further downloading the dependency ebuilds?......

this makes the issue with the serch so lets transfer the idea of the textfile over to a db with ebuildname and all the other info that comes up in emerge -s?

so now that is a easy way to not have a hell of a long time doing an emerge sync (not it's realy that long but still and we also have minimised the serch time with some kind of db)???

now this makes some non broadband users yell hey what if i want to be able to do an emerge -p without having to be online well for those users we still got the ebuilds on rsync mirrors to download the tree if they want to or maybe just all the ebuilds that are used in "world" during the sync process.

P.S. if you think that i think that i've got wings or something else od let me know ;)
_________________
What? Where? Oh! this one, it's Gentoo.
If you don't have anything constructive to say, you might want to consider staying quiet instead.
To many destructive comments, which in no way will help making the Gentoo community better.
Back to top
View user's profile Send private message
dol-sen
Retired Dev
Retired Dev


Joined: 30 Jun 2002
Posts: 2805
Location: Richmond, BC, Canada

PostPosted: Fri Aug 15, 2003 6:03 am    Post subject: Reply with quote

Many of thoe features everyone is talking about have been or are being worked on, just not a single co-ordinated effort. For a database driven portage check out portage-sql being worked on at breakmygentoo.net.

As for a database driven portage tree, non-broadband users would not have a problem with dependancy checking, etc. as the database should contain all pertinent data about the package, just not the ebuild itself. The ebuild should be downloaded with the package, for non-broadband/offline emerges there would still be the
Code:
emerge -f [package list]
option.

As for better progress indicators, emerge already ouputs title bar info as said earlier. Could emerge not also spawn another terminal window and pipe more detailed info to that terminal, such as -p pkg list. Also open another window (or pipe to another file) and output all ebuild configure warnings/suggestions to that window for later reference and checking.

As for binary vs text file database , I am in favour of using xml. Also if I'm not mistaken xml also includes the ability to store binary data strings as part of the file/record If it proved to be advantageous/neccesary.

The other feature I would like to see is a database file generated/added to containing the files merged for a package so newb's and failing memory guru's can figure out what command line app it is that I need & are available when you can't find a matching manpage. Then spend a lot of time trying to find it, or re-emerge it to see the last output from emerge showing what file went where. Or is that something that only happens to me? I just remembered that info could be found in /var/db/pkg/[category]/[package]/CONTENTS . Something a database driven portage could simplify.

An added command could be
Code:
 big_squirt brian # pkghelp -f dvdrtools

could output the installed file list for dvdrtools. Possible options could be -f = all file installed, -c = only runnable commands, -m = man/info pages only

Laastly, the c/c++ issue, the last I saw on official portage development was that it was going thru a major change, portage main will be a c/c++ coded daemon, with many of emerge's current functions moved to portage daemon. But that was several months ago.
_________________
Brian
Porthole, the Portage GUI frontend irc@freenode: #gentoo-guis, #porthole, Blog
layman, gentoolkit, CoreBuilder, esearch...
Back to top
View user's profile Send private message
charlieg
Advocate
Advocate


Joined: 30 Jul 2002
Posts: 2149
Location: Manchester UK

PostPosted: Fri Aug 15, 2003 11:24 am    Post subject: Reply with quote

Senso wrote:
Tronic wrote:

Someone suggested SQL.. Dunno about its search performance either, could be fast too..


It IS extremely fast, compared to normal text searching/indexing. :) It's in binary so not human-readable but MySQL or PostgreSQL (and I hope SQLite) allow a lot of flexibility if you need to change something. Also, Python has a good MySQL support.


Why use something as big as mysql or postgres when there are much more compact options available: the Berkerley DB and Metakit.
_________________
Want Free games?
Free Gamer - open source games list & commentary

Open source web-enabled rich UI platform: Vexi
Back to top
View user's profile Send private message
Joe
n00b
n00b


Joined: 01 Jul 2002
Posts: 70
Location: Europe

PostPosted: Fri Aug 15, 2003 11:56 am    Post subject: USE flag nonsense Reply with quote

I am a little disappointed by the USE Flags read by emerge.

First, negative USE flags are a bad idea IMHO. Something like moznoirc, moznocompose, nocardbus etc. should be avoided.
Call them mozirc, mozcompose and cardbus and leave them unset.

Second, the system should take notice of the fact, that there are no global USE flags for your box.

Actually I have to adjust my flags for every package installation (and for each rebuild/update!).

For example, if I want to emerge php, I need postgresql but not X and pam. Later I want to emerge courier, but I dont want courier to use postgresql but pam. For postgresql I dont want SSL support (just local access required), but php should use it of course and so on.

In short words: Its kind of impossible to find a USE flag configuration which fits to your needs over all packages going to be compiled.

Yeah, ufed is your friend, but anyway I think that was not the idea behind the USE flags.
Besides, an "emerge world" is quite useless for me without "-vp".


Third, I want a possibility to find out afterwards with which flags set a package has been compiled. As you can imagine I am not able to find out which flags I used a month ago by peeking into /etc/make.conf.

Fourth and last, I would like the possibility to see a changelog or at least some quick info why I should upgrade a package without having to visit the projects homepage.

Regards,

Joacher
Back to top
View user's profile Send private message
Senso
Apprentice
Apprentice


Joined: 17 Jun 2003
Posts: 250
Location: Montreal, Quebec

PostPosted: Fri Aug 15, 2003 12:47 pm    Post subject: Reply with quote

charlieg wrote:
Senso wrote:
Tronic wrote:

Someone suggested SQL.. Dunno about its search performance either, could be fast too..


It IS extremely fast, compared to normal text searching/indexing. :) It's in binary so not human-readable but MySQL or PostgreSQL (and I hope SQLite) allow a lot of flexibility if you need to change something. Also, Python has a good MySQL support.


Why use something as big as mysql or postgres when there are much more compact options available: the Berkerley DB and Metakit.


Because, as I already said, I find eats Metakit too much RAM. I haven't played enough with DB to talk about it. I was talking about MySQL but I was thinking about SQLite, which is embedded SQL in your app. I'm pretty sure it could compete with Metakit.
Back to top
View user's profile Send private message
Hoeken
n00b
n00b


Joined: 07 Feb 2003
Posts: 7
Location: Iowa

PostPosted: Fri Aug 15, 2003 2:53 pm    Post subject: Re: Emerge - the future? Reply with quote

Quote:

Quote:
-Instead of the regular 50 lines/second build dump, display good overall progress indicators

And how would the program know how far the build process has come? Make does not provide such information.

All these things have been discussed many times in other threads.


well, theres a program out there that does that already... its called 'splat' app-admin/splat
_________________
Educate -> Enlighten -> Evolve
http://www.neuroatomik.com
Back to top
View user's profile Send private message
scrllock
Tux's lil' helper
Tux's lil' helper


Joined: 14 Oct 2002
Posts: 102

PostPosted: Fri Aug 15, 2003 5:03 pm    Post subject: Reply with quote

nixze's point is quite good. In freebsd, using cvsup to upgrade your ports, you can specify that you don't want, say the astrology and biology sections. thus time and bandwidth and diskspace are all saved.. I don't know if rsync can do this though. however, if an alternate method of syncing was made available...
Back to top
View user's profile Send private message
M104
Tux's lil' helper
Tux's lil' helper


Joined: 13 Jan 2003
Posts: 132
Location: Riverside, CA

PostPosted: Fri Aug 15, 2003 5:58 pm    Post subject: Re: USE flag nonsense Reply with quote

Joe wrote:
I am a little disappointed by the USE Flags read by emerge.

First, negative USE flags are a bad idea IMHO. Something like moznoirc, moznocompose, nocardbus etc. should be avoided.
Call them mozirc, mozcompose and cardbus and leave them unset.


I totally agree with you on this. I am always shocked at how many USE flags there are when I use ufed. Has this number increased dramatically in the past year, or am I just seing things? Perhaps USE flags that only deal with one package should not be lumped in with flags like "X" and "static"? If Linux is going to take off on the desktop, Gentoo is going to need hundreds of USE flags! Yech! Arranging USE flags in categories might help, but it needs to be done well. Just look at the way the kernel configuration works in menuconfig, for example. You want to add a feature like SCSI and you get access to the SCSI variables and flags. Nice.

In the near future, a better way might be to prompt the use for flags that are unspecified in make.conf or the USE env variable. In other words, the default value for the more specific flags would be "unspecified," not on or off. If you tried to emerge something (or it was one of the dependencies), you would be prompted at the beginning of the emerge to specify those USE flags.

Joe wrote:
Third, I want a possibility to find out afterwards with which flags set a package has been compiled. As you can imagine I am not able to find out which flags I used a month ago by peeking into /etc/make.conf.


Check your /var/db/pkg/$CATEGORY/$PKG_NAME/USE file. :D

Joe wrote:
Fourth and last, I would like the possibility to see a changelog or at least some quick info why I should upgrade a package without having to visit the projects homepage.


emerge -pl $PKG_NAME

Oh yes, and more descriptive package descriptions would be a HUGE help for those of us who are trying to decide between similar packages.

Cheers!
_________________
"Pulling together is the aim of despotism and tyranny. Free men pull in all kinds of directions."
Terry Pratchett, The Truth
Back to top
View user's profile Send private message
dol-sen
Retired Dev
Retired Dev


Joined: 30 Jun 2002
Posts: 2805
Location: Richmond, BC, Canada

PostPosted: Sat Aug 16, 2003 7:01 am    Post subject: Reply with quote

to joe

Quote:
Actually I have to adjust my flags for every package installation (and for each rebuild/update!).

For example, if I want to emerge php, I need postgresql but not X and pam. Later I want to emerge courier, but I dont want courier to use postgresql but pam. For postgresql I dont want SSL support (just local access required), but php should use it of course and so on.

In short words: Its kind of impossible to find a USE flag configuration which fits to your needs over all packages going to be compiled.


I started work on a emerge/portage wrapper that adds this functionality by recording the commandline overrides for emerging the package and can then re-use those for later emerges/updates of that package. I do not have a working program as yet, I have been far too busy this past month, and I'll be away for the next 2 months, and I don't know how much I'll be able to work on it.

Quote:
Third, I want a possibility to find out afterwards with which flags set a package has been compiled. As you can imagine I am not able to find out which flags I used a month ago by peeking into /etc/make.conf.


That info is in the /var/db/pkg/[category]/[package]/USE file. You can use the command
Code:
emerge -vep world | grep "a_flag"
to get a list of all packages in your system that have "a_flag" as an option and how it was set when emerged. I also am adding this feature into the wrapper I am working on, along with the ability to re-emerge any that have that flag as a configurable option if desired
_________________
Brian
Porthole, the Portage GUI frontend irc@freenode: #gentoo-guis, #porthole, Blog
layman, gentoolkit, CoreBuilder, esearch...
Back to top
View user's profile Send private message
Genone
Retired Dev
Retired Dev


Joined: 14 Mar 2003
Posts: 9559
Location: beyond the rim

PostPosted: Sat Aug 16, 2003 9:12 am    Post subject: Reply with quote

Ok, I've read the majority of the posts here and I think I should adress some of the issues presented here:

portage speed:
Yes, portage is not as fast as possible. All the import statements in portage.py and emerge cause a startup delay (about 3-5s on my P2 233). The actual runtime performance is more affected by the disk accesses, not the implementation language. This is not so much an issue when emerging a package, but really annoying for searches or pretending. There are several scripts to improve search speed by using an index file (search for esearch or fastsearch).

(dis)advantages of C/C++ over python:
The speed improvements would not be that much. A major improvement would be error checking on compile time. On the other hand it's much easier to fix bugs or implement new features in python. And anyone who thinks that the ebuild syntax is tied to portage written in python is wrong: ebuilds are written in bash :)

DB backend:
I'll quote carpaski (the current portage maintainer) from an IRC discussion today (a bit edited):
carpaski wrote:
<carpaski> Look at python's "anydb" module.
<carpaski> Making a portage db frontend is a high preference. Supporting _any_ database as a pluggable module is something we should have.
<carpaski> But dbm is a good start.
<carpaski> I'd personally like to look at edb.
<dma`> I see anydbm, bsddb, but no anydb
<nakano> carpaski: yep. i think so. we should make a interface for db.
<carpaski> dma`: That's it.
<dma`> will this be compatible with the existing dbapis?
<carpaski> dma`: anydbm can do most 'dbm' databases.
<nakano> carpaski: are you going to implement for db?
<carpaski> Yeah, in the not-to-distant future.
<carpaski> We're going to be working toward that after 2.0.52, I'd guess.
<dma`> and how many .db/.dbm files would be used?
<carpaski> One for /var/cache/edb/
<carpaski> _possibly_ one for /var/db/pkg, but not likely. I want that to be flat text.
<carpaski> Extentions (langs, and other random things) could go in other ones.

So this will be implemented in one of the next versions (might still take some months), but not for the main tree.

USE flag / KEYWORD issues:
there is a working patch on bug 13616 to get a package.use and package.keywords file that will override global settings. The experimental version in my sig also includes that patch.

Threaded portage:
This is an old issue and yes, it's problematic AFAIK.

Security updates:
I've already written some code for this, but it needs support from the GLSA guys. Search the archives of the -dev, -server and -security mailing lists (at least -dev has an archive), there were multiple discussions about this.

Downloading all ebuilds:
cat /etc/make.conf wrote:

# RSYNC_EXCLUDEFROM is a file that portage will pass to rsync when it updates
# the portage tree. Specific chucks of the tree may be excluded from
# consideration. This may cause dependency failures if you are not careful.
# The file format is one pattern per line, blanks and ';' or '#' lines are
# comments. See 'man rsync' for more details on the exclude-from format.
#RSYNC_EXCLUDEFROM=/etc/portage/rsync_excludes

I exclude several categories this way.

I hope I could bring some light to most of your issues. I know that portage is very far from being perfect, but most of the issues are known and being worked on.
Please note that I'm not officially working for Gentoo so this is only my personal knowledge that I gathered from bugzilla, mailing lists and IRC and by no means an official statement.
Back to top
View user's profile Send private message
Lowspirit
Apprentice
Apprentice


Joined: 31 Jul 2002
Posts: 258
Location: Northern Sweden

PostPosted: Sat Aug 16, 2003 9:43 am    Post subject: Reply with quote

Let's pretend I'm doing a big multi-ebuild emerge over ssh, how do I know which package it is on then and of how many/etc? I tend to login with a new user and just "emerge package -p" and see how many is left but that's not smoooth.

All suggestions in this thread for that "flaw" was aimed towards X usage as far as I could tell.
_________________
Gentoo | AMD X2 3800+ 2GB RAM | Kernel 2.6.30 . ReiserFS . CFQ . GCC4.3.3 | Firefox 3.5 | Gnome 2.26 w/ Compiz-Fusion
"Penguins are the only fish that can fly"
Back to top
View user's profile Send private message
Joe
n00b
n00b


Joined: 01 Jul 2002
Posts: 70
Location: Europe

PostPosted: Sat Aug 16, 2003 10:59 am    Post subject: Re: USE flag nonsense Reply with quote

Quote:
I totally agree with you on this. I am always shocked at how many USE flags there are when I use ufed. [...]
Arranging USE flags in categories might help, but it needs to be done well. Just look at the way the kernel configuration works in menuconfig, for example.


Yes, that would be the very best solution. The possibility to store the USE flags for each specific package.


Thanks for the help with my other problems!

Regards,

Joe
Back to top
View user's profile Send private message
plate
Bodhisattva
Bodhisattva


Joined: 25 Jul 2002
Posts: 1663
Location: Berlin

PostPosted: Sat Aug 16, 2003 12:13 pm    Post subject: Reply with quote

Lowspirit wrote:
Let's pretend I'm doing a big multi-ebuild emerge over ssh, how do I know which package it is on then and of how many/etc?

Open another ssh session and do
Code:
tail -f /var/log/emerge.log
Back to top
View user's profile Send private message
kernja
n00b
n00b


Joined: 12 Jun 2003
Posts: 65

PostPosted: Sun Aug 17, 2003 1:44 am    Post subject: Reply with quote

What I'd like to see is when doing an "emerge -u world", for example, if one app fails to compile the emerge won't stop but will just not try to compile anything dependent on the failed package. I'm sure most of us have gone to sleep after running"emerge -u world" only to find it the morning that two packages were compiled the third failed and 50 others were aborted.
Back to top
View user's profile Send private message
Ox-
Guru
Guru


Joined: 19 Jun 2003
Posts: 305

PostPosted: Sun Aug 17, 2003 5:59 am    Post subject: Reply with quote

Genone wrote:
A major improvement would be error checking on compile time. On the other hand it's much easier to fix bugs or implement new features in python.

The benefits of static typing checked at compile time (e.g. C++) vs. dynamic typing checked at runtime (e.g. Python) is something that is argued regularly, but the arguments are usually related to bugs. So, I'm curious as to what you mean by "major improvement"?
Back to top
View user's profile Send private message
Genone
Retired Dev
Retired Dev


Joined: 14 Mar 2003
Posts: 9559
Location: beyond the rim

PostPosted: Sun Aug 17, 2003 2:19 pm    Post subject: Reply with quote

Ox- wrote:
Genone wrote:
A major improvement would be error checking on compile time. On the other hand it's much easier to fix bugs or implement new features in python.

The benefits of static typing checked at compile time (e.g. C++) vs. dynamic typing checked at runtime (e.g. Python) is something that is argued regularly, but the arguments are usually related to bugs. So, I'm curious as to what you mean by "major improvement"?


I meant that bugs would be detected earlier and mistakes like the -r4 and -r6 releases won't happen again.
Back to top
View user's profile Send private message
Ox-
Guru
Guru


Joined: 19 Jun 2003
Posts: 305

PostPosted: Mon Aug 18, 2003 8:42 am    Post subject: Reply with quote

Genone wrote:
I meant that bugs would be detected earlier and mistakes like the -r4 and -r6 releases won't happen again.

Ah, ok :lol:

Yes, it is a major improvement to get working software over non-working software, but IMHO I think the fault is more with the programmer than the language or even a specific implementation of a language.

Btw, I don't know if you've seen any of the serious discussions along these lines so I apologize if you have, but check out what these guys have to say: Bruce Eckel (a guy who makes his living writing books and speaking at conferences about Java and C++) and Robert Martin (another C++ author).
Back to top
View user's profile Send private message
tactless
l33t
l33t


Joined: 14 Jul 2002
Posts: 642
Location: Mitzpe Adi, Israel

PostPosted: Mon Aug 18, 2003 12:45 pm    Post subject: Reply with quote

A little note about the progress meters: SBUs, or Static Bash Units. Time a static compilation of bash, and, for example, a compilation of the latest xfree86. Divide, and you get the compilation time of xfree86 in SBUs. Add this information to the ebuild.
Then, have the user do something like "emerge sbtime" to figure out how long an SBU takes on his own system. When he emerges xfree86, it'll tell him approximately how long that will take on his system, running under the same parameters as the static bash compilation. Not precise, but better than what we have now.
_________________
Tactless

"If it wasn't for fog, the world would run at a really crappy framerate."

Jabber: tactless@amessage.info
Back to top
View user's profile Send private message
Ximinez
n00b
n00b


Joined: 28 Apr 2002
Posts: 7
Location: Cork, Ireland

PostPosted: Mon Aug 18, 2003 1:06 pm    Post subject: Reply with quote

plate wrote:
Lowspirit wrote:
Let's pretend I'm doing a big multi-ebuild emerge over ssh, how do I know which package it is on then and of how many/etc?

Open another ssh session and do
Code:
tail -f /var/log/emerge.log


Another possibility is to open another ssh-session, and do a
Code:
ps aux
to see the name of the merge (It's in square brackets somewhere), and then you can check how many packages are left by doing an
Code:
emerge -p whatever you merged

Just my 0.02$
_________________
"linux" is a very common word and was not included in your search. [details]
Back to top
View user's profile Send private message
Souperman
Guru
Guru


Joined: 14 Jul 2003
Posts: 449
Location: Cape Town, South Africa

PostPosted: Mon Aug 18, 2003 1:52 pm    Post subject: Reply with quote

You should be able to see which package is currently being merged and how many are being merged in the current operation, e.g.:

Code:

root@gandalf:/root# emerge -u world
Calculating world dependencies ...done!
>>> emerge (1 of 1) sys-libs/db-3.2.9-r7 to /


In this case only 1 package was being merged, but you get the point.
_________________
moo
Back to top
View user's profile Send private message
Ximinez
n00b
n00b


Joined: 28 Apr 2002
Posts: 7
Location: Cork, Ireland

PostPosted: Mon Aug 18, 2003 2:03 pm    Post subject: Reply with quote

gpd wrote:
You should be able to see which package is currently being merged and how many are being merged in the current operation, e.g.:

Code:

root@gandalf:/root# emerge -u world
Calculating world dependencies ...done!
>>> emerge (1 of 1) sys-libs/db-3.2.9-r7 to /


In this case only 1 package was being merged, but you get the point.

Yeah, but that doesn't show which package is currently being merged, and how far it is in the process.
_________________
"linux" is a very common word and was not included in your search. [details]
Back to top
View user's profile Send private message
STEDevil
Apprentice
Apprentice


Joined: 24 Apr 2003
Posts: 156

PostPosted: Mon Aug 18, 2003 5:07 pm    Post subject: Reply with quote

Ox- wrote:
Lovechild wrote:
psstt...

www.zynot.org - the future of Portage :)

http://www.joelonsoftware.com/articles/fog0000000069.html


2 things

1) That link has nothing to do with Zynot. Zynot is not starting from scratch but building ontop of Gentoo. And while many say that there is no point in making Portage into C/C++, that is probably correct, as long as your scope is Desktops and Servers. Once you start looking into eg handhelds things change a bit.

2) I first read that link many years ago, thought he could be right in principle in the short term, but was quite sure he was wrong in the long run. Looking at today I can't recall seeing any review in the last year or so where someone has _not_ said that gecko blows MS IE away in basicly every area there is.
NS 4 was a really crappy browser with patched in CSS through JS. Sure they could have kept patching in a few new CSS things here and there, but to get truly good CSS support you need it integrated at the core which in turn means a total rewrite of the core sooner or later. As a webdeveloper I sure apreciate it happening sooner rather then later...
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3, 4, 5  Next
Page 2 of 5

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum