[SOLVED] Using lazbuild to build LazMapViewer

[MarkMLl]

A few months ago I found myself putting a bit of time into a script to build the Lazarus IDE with preinstalled packages, after finding that "using the IDE to add packages to itself" was having difficulty building for multiple widget sets (I like having precompiled IDEs for GTK and Qt) https://forum.lazarus.freepascal.org/index.php/topic,65706.msg507514.html#msg507514 .

I've been looking at that again today with a view to using it for an IDE with LazMapViewer preinstalled, and hit what I believe is a known problem relating to ssl_openssl.pas (https://forum.lazarus.freepascal.org/index.php/topic,47082.msg336821.html#msg336821 https://forum.lazarus.freepascal.org/index.php/topic,47082.msg336838.html#msg336838).

I know that my system is able to compile and run the standard LazMapViewer demos in an IDE where the packages have been installed using OPM. However I'd be very interested if anybody could recommend a "best practice" solution for setting it up using lazbuild.

I attach my current lazbuild script in case it is of any use or interest to anybody.

MarkMLl

[MarkMLl]

I've followed the lead of OLPM and created symlinks for ssl_openssl.pas, ssl_openssl_lib.pas and jedi.inc in the synapse_downloadengine directory.

Currently having problems building for Qt5 and unused packages and links have to be removed using the IDE, but I think it's good enough since it allows me to regenerate my working IDE etc. in a controlled fashion.

The form of --add-package-link in the --help output etc. appears to be wrong: the .lpk shouldn't be supplied as a parameter to the option but as the final parameter that the program is working on.

Updated shell script attached.

MarkMLl

[wp]

Sorry can't help you with this. All the make tools always have been a mystery to me (just followed by regular expressions...)

[MarkMLl]

Sorry can't help you with this. All the make tools always have been a mystery to me (just followed by regular expressions...)

OTOH, your two cited ML postings (and looking what OLPM actually did) were seminal.

MarkMLl