[SOLVED] Patching dwm the Gentoo Way

[alecStewart1]

Hello, friends!

I thought I'd give dwm a try and I like how I can put patches in /etc/portage/patches/x11-wm/dwm for patches!

I'm running into an issue installing dwm and applying the vanitygaps patch.

If you look at line 212 and 1671 in dwm.c, the vanitygaps patch removes those from dwm.c and adds them into it's vanitygaps.c.

However emerge doesn't seem to like those hunks in that patch and will fail at both of those hunks, failing to patch and install dwm.

So...what exactly do I do to get around that? I know I could fork dwm into my own git repo, add the patches and create my own local ebuild pointing to that git repo, but I'd rather try and do it the /etc/portage/patches way.

[fedeliallalinea]

I tried dwm-vanitygaps-6.2.diff, the problem is that patch is created for 6.2 version and gentoo tree as 6.3 and 6.4.
Try this patch, it works for 6.3 version.

[alecStewart1]

Try this patch, it works for 6.3 version.

Thanks! That does work and dwm compiles fine. However, is this going to be an issue for other patches as well?

In total, I'm wanting to use the

• statuscmd (to use dwmblocks)
  alt tab
  fullscreen
  shift tools
  xresources
  vanitygaps
  steam
  stacker
  sticky
  swallow

patches. On the suckless website for dwm, it looks like most of those have support of 6.2.

Is there going to be a lot of tinker with just patches before I can tinker with dwm?

[NeddySeagoon]

alecStewart1,

Try the patches and see.
Portage will tell you if they fail to apply.

[fedeliallalinea]

Thanks! That does work and dwm compiles fine. However, is this going to be an issue for other patches as well?
...
Is there going to be a lot of tinker with just patches before I can tinker with dwm?

Probably yes, but you should test it.
That's one of the reasons I chose a wm that had what I needed without having to waste time with patches.

[alecStewart1]

Thanks! That does work and dwm compiles fine. However, is this going to be an issue for other patches as well?
...
Is there going to be a lot of tinker with just patches before I can tinker with dwm?

Probably yes, but you should test it.
That's one of the reasons I chose a wm that had what I needed without having to waste time with patches.

Fair enough, but I think I've found one of those things that I stubbornly want to get to work and refuse to give up, lol.

Try the patches and see.
Portage will tell you if they fail to apply.

I'm finding that sometimes it's not hard to update patches and a total pain other times. I'm running into this problem now with the shift-tools patch.

Code: Select all

 * Applying 06-shift-tools.diff ...
patching file config.def.h
patch: **** malformed patch at line 29:  };
                                                                                                                                                                                                                                                       [ !! ]
 * ERROR: x11-wm/dwm-6.3::gentoo failed (prepare phase):
 *   patch -p1  failed with /etc/portage/patches/x11-wm/dwm/06-shift-tools.diff

Okay, so here's the hunk it's griping about:

Code: Select all

--- a/config.def.h	2022-11-07 15:48:11.352474594 -0600
+++ b/config.def.h	2022-11-07 15:52:52.046649301 -0600
@@ -93,2 +93,2 @@
 		{ "mfact",      	 	FLOAT,   &mfact },
 };

+#include "shift-tools.c"
+
 static Key keys[] = {
 	/* modifier                     key        function        argument */
 	{ MODKEY,                       XK_p,      spawn,          {.v = dmenucmd } },
 	{ MODKEY|ShiftMask,             XK_Return, spawn,          {.v = termcmd } },
+	{ MODKEY,                       XK_o, shiftviewclients,    { .i = +1 } },
+	{ MODKEY|ShiftMask,             XK_o,	shiftview,         { .i = +1 } },
+	{ MODKEY|ShiftMask,             XK_i,	shiftview,         { .i = -1 } },
+	{ MODKEY,	                XK_i, shiftviewclients,    { .i = -1 } },
 	{ MODKEY,                       XK_b,      togglebar,      {0} },
 	{ MODKEY,                       XK_j,      focusstack,     {.i = +1 } },
 	{ MODKEY,                       XK_k,      focusstack,     {.i = -1 } },
@@ -104,6 +110,10 @@
 	{ MODKEY,                       XK_d,      incnmaster,     {.i = -1 } },
 	{ MODKEY,                       XK_h,      setmfact,       {.f = -0.05} },
 	{ MODKEY,                       XK_l,      setmfact,       {.f = +0.05} },
+	{ MODKEY|ShiftMask,		XK_h,      shiftboth,      { .i = -1 }	},
+	{ MODKEY|ControlMask,		XK_h,      shiftswaptags,  { .i = -1 }	},
+	{ MODKEY|ControlMask,		XK_l,      shiftswaptags,  { .i = +1 }	},
+	{ MODKEY|ShiftMask,             XK_l,      shiftboth,      { .i = +1 }	},
 	{ MODKEY,                       XK_Return, zoom,           {0} },
 	{ MODKEY,                       XK_q,	   view,           {0} },
 	{ MODKEY|ShiftMask,             XK_c,      killclient,     {0} },

It'd be nice if patch could give a little more hints as to what makes malformed. Here's the full patch file.

[NeddySeagoon]

alecStewart1,

This looks wrong to me.

Code: Select all

@@ -93,2 +93,2 @@
       { "mfact",             FLOAT,   &mfact },
 };

+#include "shift-tools.c"
+
 static Key keys[] = {
    /* modifier                     key        function        argument */
    { MODKEY,                       XK_p,      spawn,          {.v = dmenucmd } },
    { MODKEY|ShiftMask,             XK_Return, spawn,          {.v = termcmd } },
+   { MODKEY,                       XK_o, shiftviewclients,    { .i = +1 } },
+   { MODKEY|ShiftMask,             XK_o,   shiftview,         { .i = +1 } },
+   { MODKEY|ShiftMask,             XK_i,   shiftview,         { .i = -1 } },
+   { MODKEY,                   XK_i, shiftviewclients,    { .i = -1 } },
    { MODKEY,                       XK_b,      togglebar,      {0} },
    { MODKEY,                       XK_j,      focusstack,     {.i = +1 } },
    { MODKEY,                       XK_k,      focusstack,     {.i = -1 } }, 

Code: Select all

@@ -93,2 +93,2 @@

tells patch that the hunk does not change the line count but it actually adds 6 lines. The lines prefixed with a +. There are no lines removed. They would have a - prefix.
The before and after line counts do not match what the patch actually does to the code.

Code: Select all

The next hunk @@ -104,6 +110,10 @@
 	{ MODKEY,                       XK_d,      incnmaster,     {.i = -1 } },
 	{ MODKEY,                       XK_h,      setmfact,       {.f = -0.05} },
 	{ MODKEY,                       XK_l,      setmfact,       {.f = +0.05} },
+	{ MODKEY|ShiftMask,		XK_h,      shiftboth,      { .i = -1 }	},
+	{ MODKEY|ControlMask,		XK_h,      shiftswaptags,  { .i = -1 }	},
+	{ MODKEY|ControlMask,		XK_l,      shiftswaptags,  { .i = +1 }	},
+	{ MODKEY|ShiftMask,             XK_l,      shiftboth,      { .i = +1 }	},
 	{ MODKEY,                       XK_Return, zoom,           {0} },
 	{ MODKEY,                       XK_q,	   view,           {0} },
 	{ MODKEY|ShiftMask,             XK_c,      killclient,     {0} },

is a confirmation.

Code: Select all

@@ -104,6 +110,10 @@

six lines at line 104 in the old file becomes 10 lines at line line 110 in the new file.
Looking at the shifs in start is that second hunk, its 6 lines. That will be the 6 lines added by the hunk above.
This hunk adds four lines hence the 6 becomes 10.

The

Code: Select all

@@ -93,2 +93,2 @@

is not consistent with what the hunk actually does.

Apply the changes by hand and remake the patch.

[alecStewart1]

Thanks guys!

I haven't been able to tinker away much at some of the patches, but I'm making progress.

Just as a note, fedeliallalinea, that patch for vanitygaps seems to work for 6.4 as well. I'll try to remember to post a link to a repo of the updated patches here, in case anyone else comes across this in in the future.

[fedeliallalinea]

Just as a note, fedeliallalinea, that patch for vanitygaps seems to work for 6.4 as well. I'll try to remember to post a link to a repo of the updated patches here, in case anyone else comes across this in in the future.

This is a good idea!

[Hammett]

If you want to apply several patches, some might fail because a previous patch has modified the file current patch is trying to modify...

I have gone through this myself and the approach I took was to download dwm source and apply the patches manually, fixing the errors on each patch. Then I renamed the patches to be applied in the correct order and moved those to /etc/portage/patches

If you're curious, I made a git repo for my dwm build... https://github.com/DasHammett/dwm
Not sure if it would build, but you will get an idea of how I did it.

[jb1277976_]

If you want to apply several patches, some might fail because a previous patch has modified the file current patch is trying to modify...

I have gone through this myself and the approach I took was to download dwm source and apply the patches manually, fixing the errors on each patch. Then I renamed the patches to be applied in the correct order and moved those to /etc/portage/patches

If you're curious, I made a git repo for my dwm build... https://github.com/DasHammett/dwm
Not sure if it would build, but you will get an idea of how I did it.

Hi,

I'm having this same error i apply a patch i know works and its fine. but when i try another patch it fails. i'm wondering if i have to keep the previous patch in place and the emerge ? if i download the source and patch its easy i just copy paste and re-compile and everything works. What would you do in this situation ?

Thanks

~ Joe B

[GDH-gentoo]

I'm still amazed by the fact that nobody seems to want to run unpatched DWM. Does it suck so bad (pun intended)?

The problem with the patches published at the project's website is that they are diffs against a specific upstream version. And that only really works flawlessly if you want to apply only one patch, to that unmodified specific version [1]. Makes sense for a quick bug fix. But doing what amounts to a nontrivial rewrite of the original software through a series of patches? Therefore, what I said in the first paragraph.

[1] So the second and later patches in a sequence are no longer applied to the unmodified version, but to code already modified by earlier patches. That only works by chance if you are lucky.

[Hu]

I'm having this same error i apply a patch i know works and its fine. but when i try another patch it fails.

Without more detail, it is hard for us to determine whether you did this correctly.

i'm wondering if i have to keep the previous patch in place and the emerge ?

Yes, you need to keep active all the patches that you want to be active.

if i download the source and patch its easy i just copy paste and re-compile and everything works. What would you do in this situation ?

I would use a project that has mastered the use of a configure system to enable and disable features via a dedicated configure phase, without resorting to source code patches for every commonly used feature.

[jb1277976_]

I'm having this same error i apply a patch i know works and its fine. but when i try another patch it fails.

Without more detail, it is hard for us to determine whether you did this correctly.

i'm wondering if i have to keep the previous patch in place and the emerge ?

Yes, you need to keep active all the patches that you want to be active.

if i download the source and patch its easy i just copy paste and re-compile and everything works. What would you do in this situation ?

I would use a project that has mastered the use of a configure system to enable and disable features via a dedicated configure phase, without resorting to source code patches for every commonly used feature.

When downloading dwm directly from suckless and apply patchces everything works fine.
I have to copy and paste code here and there but everything is working. I just wanted to try it the gentoo way since i'm on gentoo now.

Harder then it looks

This looks about right viewtopic-p-8815183.html#8815183

[NeddySeagoon]

Moved from Desktop Environments to Portage & Programming.