Commits on Sep 23, 2022

1. font-patcher: Remove "SrcEnd" option …

   [why]
   In patch set definitions we have for the source ranges SymStart and SymEnd
   and for the destination we can specify SrcStart and SrcEnd
   
   The SrcEnd can be automatically generated. For SrcEnd values that differ
   from the autogenerated value (are lower) the script would crash, or (are
   higher) ignore them anyhow.
   
   There are two modes: 'exact = True' and 'exact = False'.
   
   The SrcStart and SrcEnd values are ignored if exact is True, because the
   glyphs are patched into the same codepoint where they originate. This
   also means that gaps in the symbols are preserved - all patched in
   glyphs have the same codepoint as they have in the source (symbol) font.
   
   When exact is False on the other hand, all (non empty) glyphs are filled
   into the codepoints that start at SrcStart. Gaps (empty glyphs) are
   ignored and thus are not present as gap in the patched font anymore (*).
   The to-be-filled-next codepoint in the patched font just increases by 1
   on every filled symbol.
   
   See note for the reason.
   
   This also makes maintining the patch set easier as noone needs to
   'calculate' a SrcEnd value anymore.
   
   [how]
   Use directly the start value and the counter instead of filling an array
   with a range, that is then indexed by the counter.
   
   Before this commit:
   
     list_of_patched_font_codepoints = list(range(SrcStart, SrcEnd + 1))
     current_codepoint = list_of_patched_font_codepoints[loop_counter]
   
   After this commit:
   
     current_codepoint = SrcStart + loop_counter
   
   [note]
   Maybe related to code removed with c728079.
   
   I guess the code uses the list, because it has been believed that glyphs
   can only be indexed by strings containing a hex number. The array
   supposedly contained that strings.
   
   But in fact the fontforge python module docu is like this:
   
   font.__getitem__(key)
   
       If key is an integer, then returns the glyph at that encoding.
       If a string then returns the glyph with that name.
       May not be assigned to.
   
   Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>

   

   Finii committed Sep 23, 2022
   Configuration menu

   • View commit details
   • Copy full SHA for 7ca44d0
   • Browse repository at this point

   Copy the full SHA

   7ca44d0 View commit details

   Browse the repository at this point in the history

2. font-patcher: Make symbol codepoints (more) stable …

   [why]
   The Font Logos and Octicons codepoints depend on the parallel existence
   of FontAwesome. I.e. Font Logos is shifted of Octicons or FontAwesome is
   also present in the patched font; Octicons is shifted if FontAwesome is
   present.
   
   This means that people, although using a Nerd Font, can expect the
   symbols in different locations. The reason is clear; people that just
   want one or some symbols and use a specifically patched font will be
   able to use the original symbol font codepoints.
   
   But I guess that these uses are nonexisting. Almost all will use
   'complete' patched fonts and that is what we deliver and document.
   
   To make the documentation less complicated we should fix the code point
   ranges that a specific symbol set will be patched in at.
   
   [how]
   Just drop the associated functions and use a False constant instead.
   
   [note]
   The two possible places where Octicons / Font Logos ends up are there
   since they have been added back in 2015/6 (commits 9620d47, f933b5a).
   
   Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>

   

   Finii committed Sep 23, 2022
   Configuration menu

   • View commit details
   • Copy full SHA for 92d8a60
   • Browse repository at this point

   Copy the full SHA

   92d8a60 View commit details

   Browse the repository at this point in the history

3. font-patcher: Clean up patch sets …

   [why]
   The patch set table has 'contradicting' or 'complicated' entries.
   
   [how]
   When we use exact patching a SrcStart will be ignored and shall be None
   to make that clear.
   
   The other two cases patch in only one glyph, make the entries more easy
   we could either make them 'exact' (reuse source codepoint) or specify a
   SrcStart. At the moment they rely on the (hidden?) rule that non-exact
   entries without SrcStart still reuse the SymStart...
   
   Signed-off-by: Fini Jastrow <ulf.fini.jastrow@desy.de>

   

   Finii committed Sep 23, 2022
   Configuration menu

   • View commit details
   • Copy full SHA for 501dd29
   • Browse repository at this point

   Copy the full SHA

   501dd29 View commit details

   Browse the repository at this point in the history