Re: good patch vs. sed fixes

在 2024/11/21 12:45, Bruce Dubbs ([email protected] via 
blfs-dev Mailing List) 写道:
> On 11/20/24 20:40, Uwe Düffert ([email protected] via blfs-dev 
> Mailing List) wrote:
>> Hi all,
>>
>> I'm not sure it is worth the effort, but from time to time I stumble 
>> across the "problem" that certain sed or patch commands would apply 
>> multiple times although they should not. I'm fully aware this is 
>> unsupported "deviating from the book" but it happens. I'm trying to 
>> build some not-yet-in-the-book new version, something fails, I fix the 
>> problem and try again, then some unexpected "double sed" or "double 
>> patch" silently breaks stuff. Sure, I should have done it cleanly from 
>> scratch, but the nature of the fixed problem suggested I could just 
>> repeat the build...
>>
>> Current example: I was upgrading to icu-76.1 . icu is an ugly package 
>> that usually triggers having to rebuild quite a number of others. 
>> firefox is one of those, at least when using system-icu. I'm aware the 
>> book stays with firefox ESR versions for a reason, however, if I'm 
>> rebuilding it anyway, I (try to) use the most recent release version. 
>> The ICU74+ sed for firefox in the book fixes the build problem just 
>> fine, BUT it would apply arbitrarily often. Took me quite some time to 
>> understand that my firefox-132 problem is NOT icu-76.1 but applying 
>> the system-icu sed multiple times.
>>
>> My personal solution for that is protecting my build script like this:
>> grep -q "CLASS_CHARACTER,CLASS_CHARACTER" intl/lwbrk/LineBreaker.cpp 
>> || for i in {43..47}; do
>>     sed '/ZWJ/s/}/,CLASS_CHARACTER&/' -i intl/lwbrk/LineBreaker.cpp
>> done
>>
>> I do like the idea to fix stuff with a simple sed if it helps avoiding 
>> a patch, BUT only if it does the job similarily well. I kind of 
>> appreciate the educational value of a sophisticated sed, but in my 
>> experience a unified patch is less likely to "doubly apply" (though it 
>> happened to me before) and sometimes makes clearer what actually needs 
>> to be changed.
>>
>> I'm not proposing using some more official patch such as https:// 
>> gitweb.gentoo.org/ repo/gentoo.git/commit/? 
>> id=d9712b1358dc1b64933c3090d72153aae3270f25 in this case, I'd just 
>> like to mention that this patch 1) makes the idea more clear (there 
>> are not just 5 more CLASS_CHARACTERs but they do have names) and it 
>> does not apply multiple times...
> 
> To avoid the problem, add this to your scripts
> 
> DIR=`pwd`
> rm -rf <build location>
> mkdir  <build location>
> cd     <build location>
> tar -xf $DIR/<tarball>
> cd <extracted directory>
> 
> continue with build instructions.
> 
> That way you are always working from a clean directory and any seds or 
> patches are only applied as specified in the book.
> 
> Note that the primary advantage of our use of seds in BLFS is to allow 
> you to see simple changes without downloading an extra file and looking 
> inside it to see what it is doing.  Of course more complicated changes 
> still require a patch.
> 
>    -- Bruce
> 
> 
JFYI: It will save some time to use ccache and sccache with this 
approach, since the sed typically modify only a few files. To use them 
as early as possible, I'm using simple scripts to statically build them 
on the host system, then copy them to the target LFS system and do some 
configurations.

-- 
http://lists.linuxfromscratch.org/sympa/info/blfs-dev
Unsubscribe: See the above information page