[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Fwd: Need a Windows binary of the latest makeinfo for a test
From: |
Eli Zaretskii |
Subject: |
Re: Fwd: Need a Windows binary of the latest makeinfo for a test |
Date: |
Wed, 29 Feb 2012 20:44:26 +0200 |
> Date: Wed, 29 Feb 2012 09:24:26 -0800
> From: Mark Messer <address@hidden>
> Cc: Philip Nienhuis <address@hidden>, Jordi GutiƩrrez Hermoso
> <address@hidden>,
> Tatsuro MATSUOKA <address@hidden>, Nitzan Arazi <address@hidden>,
> "a.dawid" <address@hidden>, address@hidden
>
>
> At the GNU Octave project, there is an intermittent bug that only affects
> Windows installations. 'help foo' is supposed to display information about
> the foo command. It does, but the text is garbled on some computers. See
> https://savannah.gnu.org/bugs/?35187
>
> The problem was traced to the most recent release of makeinfo.exe, which is
> 4.13 from 2008. The most recent snapshot of makeinfo.exe fixes the problem.
I believe by "the most recent snapshot" you mean the precompiled
binary from this site:
http://sourceforge.net/projects/ezwinports/files/texinfo-4.13a-w32-bin.zip/download
If so, then this is not a development snapshot, it is more or less
straightforward MinGW build of stock Texinfo 4.13a distributed by the
upstream Texinfo maintainers. (In fact, the C implementation of
makeinfo is no longer developed, and instead a new implementation in
Perl is being developed as we speak.)
> The Octave project would like to include that latest snapshot of
> makeinfo.exe in their next release.
Feel free to use the precompiled binary from the above site.
> makeinfo --version reports 4.13 for both the 2008 released version and the
> snapshot. It would be nice to have a snapshot with a different version. Is
> this possible?
The version string comes from the source code, which I didn't change.
The fact that other binaries of the same version crash, while this one
doesn't, is because I built this binary and carefully tested it to
make sure it works. I guess the binaries that crash were linked with
some faulty DLLs and most likely saw much less testing. But no
source-level differences exist between these binaries.
Why do you care about the version, anyway? I'm reluctant to change
the version string without any source-level changes in the code.