| 1 |
Guard against including a ragel file more than once (newsgroup). |
| 2 |
|
| 3 |
Line numbers in included files refer to the master file. This needs to be |
| 4 |
fixed -- from Manoj. |
| 5 |
|
| 6 |
Remove old action embedding and condition setting syntax. |
| 7 |
|
| 8 |
fbreak should advance the current char. Depreciate fbreak and add |
| 9 |
fctl_break; |
| 10 |
fctl_return <expr>; |
| 11 |
fctl_goto <label>; |
| 12 |
This is needed for better support of pull scanners. |
| 13 |
|
| 14 |
Eliminate tokend, replace it with the marker variable and add ftokend/ftoklen. |
| 15 |
|
| 16 |
Add the accept action embedding operator: like eof but only for states that end |
| 17 |
up final. Add the combined pending/accept action. This becomes a good idea when |
| 18 |
eof action execution is moved into the main loop. |
| 19 |
|
| 20 |
Add a prefix operator which sets every state final. |
| 21 |
|
| 22 |
Minimization should remove a condition when the character allows both |
| 23 |
the positive and negative sense of the condition. This happens in: |
| 24 |
test_every_10_chars = ( ( c when test_len ) c{0,9} )**; |
| 25 |
In this example there is non-determinsm that is killed by the priorities, but |
| 26 |
since conditions are expanded before priorities are tested, many transitions |
| 27 |
end up with test_len || !test_len. |
| 28 |
|
| 29 |
Should be possible to include scanner definitions in another scanner. |
| 30 |
|
| 31 |
Need an "entry name;" feature, causing name to get written out with the other |
| 32 |
entry points in the data. |
| 33 |
|
| 34 |
Possibly bring back the old semantics of > in a new operator, or allow it to be |
| 35 |
defined somehow. |
| 36 |
|
| 37 |
When priorities are embedded without a name, the name of the current machine is |
| 38 |
used. Perhaps a unique name for each instance of the current machine should be |
| 39 |
used instead. This idea could work well if applied to user-defined embeddings. |
| 40 |
|
| 41 |
User defined embeddings <-name(a1,a2,...). |
| 42 |
User defined operators expr1 <name> expr2. |
| 43 |
|
| 44 |
Doesn't make make sense for [] to be the lambda (zero-length) machine. This |
| 45 |
should be the empty set (the empty machine). But then would it be invalid in a |
| 46 |
regular expression? |
| 47 |
|
| 48 |
The |> guarded operator and the <| guarded operator need to be added. |
| 49 |
|
| 50 |
An option to turn off the removal of duplicate actions might be useful for |
| 51 |
analyzing unintentional nondeterminism. |
| 52 |
|
| 53 |
Might be a good idea to add in some protection against using up all of a |
| 54 |
system's memory. This protection could then be removed by people when someone |
| 55 |
is sure they want to use a lot of memory. |
| 56 |
|
| 57 |
|
| 58 |
If a scanner can be optimized into a pure state machine, maybe permit it to be |
| 59 |
referenced as a machine definition. Alternately: inline scanners with an |
| 60 |
explicit exit pattern. |
| 61 |
|
| 62 |
The split codegen needs a profiler connected to a graph partitioning algorithm. |
| 63 |
|
| 64 |
Die a graceful death when rlcodegen -F receives large alphabets. |
| 65 |
|
| 66 |
It's not currently possible to have more than one machine in a single function |
| 67 |
because of label conflicts. Labels should have a unique prefix. |
| 68 |
|
| 69 |
Emit a warning when a subtraction has no effect. |
| 70 |
|
| 71 |
Emit a warning when unnamed priorities are used in longest match machines. |
| 72 |
These priorities may unexpectedly interact across longest-match items. Changing |
| 73 |
the language such that unwated interaction cannot happen would require naming |
| 74 |
longest-match items. |
| 75 |
|
| 76 |
Testing facilities: Quick easy way to query which strings are accepted. |
| 77 |
Enumerate all accepted strings. From Nicholas Maxwell Lester. |
| 78 |
|
| 79 |
Add more examples, add more tests and write more documentation. |
| 80 |
|
| 81 |
A debugger would be nice. Ragel could emit a special debug version that |
| 82 |
prompted for debug commands that allowed the user to step through the machine |
| 83 |
and get details about where they are in their RL. |
| 84 |
|
| 85 |
A quick and easy alternative would be a trace code generation option. This |
| 86 |
would open a trace file and list all the active machines at each step of the |
| 87 |
input. |
| 88 |
|
| 89 |
Frontend should allow the redefinition of fsm section delimiters. |
| 90 |
|
| 91 |
Do more to obscure ragel's private variables. Just a leading underscore is not |
| 92 |
enough. Maybe something more like __ri__. |
| 93 |
|
| 94 |
Some talk about capturing data: |
| 95 |
|
| 96 |
Separate tokstart/tokend from the backtracking. One var for preservation, |
| 97 |
called preserve. Write delcarations; produces the necessary variables used by |
| 98 |
ragel. Move pattern start pattern end concepts into the general? The |
| 99 |
variables which may need to influence the preserve is dependent on the state. |
| 100 |
States have a concept of which variables are in use. Can be used for length |
| 101 |
restrictions. If there is an exit pattern, it is the explicit way out, |
| 102 |
otherwise the start state and all final states are a way out. |