Zsh Mailing List Archive
Messages sorted by:
Reverse Date,
Date,
Thread,
Author
Re: PATCH: Allow parallel make
- X-seq: zsh-workers 55142
- From: Mikael Magnusson <mikachu@xxxxxxxxx>
- To: zsh-workers@xxxxxxx
- Subject: Re: PATCH: Allow parallel make
- Date: Tue, 8 Sep 2026 22:17:07 +0200
- Arc-authentication-results: i=1; mx.google.com; arc=none
- Arc-message-signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20260327; h=content-transfer-encoding:to:subject:message-id:date:from :in-reply-to:references:mime-version:dkim-signature; bh=DQzax2CRgh+387+lHKR/HAvGXcEIcKEGxPaFMxrMEX0=; fh=SbTlPuNNxBzTkRlwWtqw/TXBY0HvGvtE97RpPp3sJPM=; b=pLSocA58iCx2OPQTJnxAeM6i2SJTeHx57xhkjRDCQQyeF/j0KXVvNZu9rILxLKD3O4 ZIECo6cJuxebFOPXyZAMTH6Kqg0ODrg+By4YpRPA6ba6P2PtZtVmTmaB9x0rlmrW2Do6 dseHiGC8i1Esb77cN3lqhUgTIF7QWhjbrfUSv14BPiuFvhVubbUZp2EHJNqY1SCInut5 ikZF5VJbQkLJp6b7qL7+oOxCxyHDdXUUH2r9dSkbzuNImNBWYcMNogKWin4XHpv60SSc tW5dYF2uGEYAgAitVtEW1A+D5466xCxAL2inWjcBUzrfv2Niaf7cdPRe1hmOzF48EsTB UBdw==; darn=zsh.org
- Arc-seal: i=1; a=rsa-sha256; t=1788898640; cv=none; d=google.com; s=arc-20260327; b=Dzt+SI9196XDAR0Gs7p8iiDKf2YBif8YmDT3RHO1F6IJFftTrKsTS/VhymesaIX+d6 KB3YIXYQzdS6mV+zCpdl1F9DanHx62JYS7egpJBQb0hYPUfx38xyhlTnLcoOaoyTEBSz pYrnXkgt9279vn/36NR5rbdickWPCXwsvvVKEKtoNI+1pw7Z1noZxGsZi6Q1SrwH2sgX sqMAxFbJ1+KHMHkMXR31quJJrtHokcC4jeOePRrrTawWFsUSAHsIVT+8pbB9pb1ScXNm etCWx25b+9k5Wr9hgA4y6wa7Q24ZoF8t6HIczcdrblsDQLiVv1ipXLtS5gr365PZp0kf /8XA==
- Archived-at: <https://zsh.org/workers/55142>
- In-reply-to: <20260908142123.5889-1-mikachu@gmail.com>
- List-id: <zsh-workers.zsh.org>
- References: <20260908142123.5889-1-mikachu@gmail.com>
On Tue, Sep 8, 2026 at 4:21 PM Mikael Magnusson <mikachu@xxxxxxxxx> wrote:
>
> ---
>
> Back in workers/23102 Vincent reported that make -j2 didn't work, and we
> added .NOTPARALLEL etc to the build rules. I finally got a bit bored
> with seeing my cpu sit at 50% when building zsh (yes, I only have a dual
> core cpu), so here's my stab at fixing it.
>
> The rule depending on headers in src/Makefile.in is the main thing that
> avoids the error Vincent saw, but there's also some follow-on errors
> that needed fixing up.
>
> I won't commit this unless a lot of people says it works on various make
> implementations (I only have GNU Make), and please try it at different
> job counts. The newuser.c failure only happened to me with make -j3, not
> make -j10 for example. The others didn't happen and are purely by
> inspection.
>
> Tighten up the atomicity of the header generation rules in zsh.mdd now
> that they could in theory be invoked concurrently.
>
> The phony rules in curses.mdd and terminfo.mdd are because I have some
> vague memory that Make may be unhappy with defining multiple identical
> rules in the same Makefile, and the mdd files' rules all end up combined
> somewhere (I assume), so I used two phony rules instead. I didn't bother
> doing that for newuser.mdd since it's the only one defining a rule for
> ../zshpaths.h
These aren't built in the default target, but they all wrote to the
same intermediate files. With --tidy they don't do that, so add that.
(Will fold this into the previous commit, if it is found to work
well).
diff --git i/Doc/Makefile.in w/Doc/Makefile.in
index 3b68ed61a1..0731e278b7 100644
--- i/Doc/Makefile.in
+++ w/Doc/Makefile.in
@@ -39,9 +39,9 @@ LN_S = @LN_S@
@DEFS_MK@
MAKEINFO = makeinfo
-TEXI2DVI = @TEXI2DVI@
+TEXI2DVI = @TEXI2DVI@ --tidy
DVIPS = dvips
-TEXI2PDF = @TEXI2PDF@
+TEXI2PDF = @TEXI2PDF@ --tidy
TEXI2HTML = @TEXI2HTML@
SET_TEXI2ANY_VAR = @SET_TEXI2ANY_VAR@
PAPERSIZE = @PAPERSIZE@
@@ -416,6 +416,7 @@ clean-here:
rm -f *.html *.info* *.dvi *.ps *.pdf
rm -f *.aux *.cp *.cps *.fn *.fns *.ky *.kys *.log *.idx
rm -f *.pg *.pgs *.toc *.tp *.tps *.vr *.vrs
+ rm -rf *.t2d
rm -rf infodir
rm -f texi2html.conf
.PHONY: clean-here
--
Mikael Magnusson
Messages sorted by:
Reverse Date,
Date,
Thread,
Author