Hi Werner,
*> Well, you say 'most'. Is this, say, 70% or 90% suitability in your> opinion? If it's the former, how much additional code must be added> to fit the needs? There might be a point that you have to add so much> code that it might be easier to avoid the library altogether.* According to me the requirements from the external library were: 1. Write logs to file (with support on different platforms including Windows). 2. Formatting of log messages according to user needs (toggle-able: timestamps and FT_COMPONENT values) These requirements are completely satisfied by dlg, and we only have to call the respective APIs given by dlg with some tweaks in FreeType to avail these features... Other than the above requirements there was a requirement to dynamically change log levels and change the output mode through the use of callback functions. For this requirement, we have to add code irrespective of whether we use an external library or we develop a logger inside FreeType. Therefore, according to me if you are concerned about the additional code that we need to add along with using dlg, it would be much less than the code if we develop a logger inside FreeType with all of the above the requirements and support on different platforms. I would try to document all the dlg's features that I am going to use along with how I am going to use them in week 4 and week 5 progress reports And that would be my plan on how I am going to integrate dlg with FreeType (obviously changeable according to the suggestions) :-) Please guide me... Thanks, Priyesh On Thu, Jun 25, 2020 at 8:24 AM Werner LEMBERG <[email protected]> wrote: > > > > The recent discussion with Behdad reminded me to ask you whether you > > > have also investigated the case where you modify the existing > > > debugging and tracing code of FreeType without using an external > > > library. For example, such new code might be based on the code from > > > 'dlg' but stripped down to the bare features we would like to see. > > > > I haven't tried this case as I thought that we need to look at this > > option only when none of the external libraries satisfies the > > specified requirements. And while exploring we have found a library > > *dlg* which is satisfying most of the requirements... > > Well, you say 'most'. Is this, say, 70% or 90% suitability in your > opinion? If it's the former, how much additional code must be added > to fit the needs? There might be a point that you have to add so much > code that it might be easier to avoid the library altogether. > > Please note that I'm only looking at the problem from different points > of view. If you say 'dlg' is the right choice to proceed, then by all > means let's do it :-) > > > Werner >
